Re: [Enigma-devel] Maemo Port for Nokia Internet Tablets

[Ronald Lamprecht]

Hi,

pipeline:
> > The upcoming Enigma 1.10 release is much faster than 1.0*. Especially levels 
> > with many actors will speedup by a factor of 5 to 10! I would expect nearly all 
> > levels to run on current cpu's without any code changes.
>
> As I mentioned above I am more than happy with the speed!  This does in fact 
> seem to make 100% of the levels playable.  This will require extensive 
> playtesting which i am more than happy to help out with :)

That is good news :-)

> Is there any timeframe in mind for when an official 1.10 release will occur? 
> Also is it considered bad practice for downstream developers to release builds 
> based on svn developmental builds?  If this varies by project, what is enigma 
> dev policy on this?  Since the performance improvements are so great it will be 
> tempting to get it out in the 'wild'.

Half a year ago I did estimate the scheduled work on 1.10 to take until 
end of March. Unfortunatly the feedback and testing support on the new 
api has been less than expected and the urgent demand on internal engine 
reengineering has been proven to be manifold. But meanwhile I finished 
the complex and critical parts without loss of compatibility or 
stability. Thus I would estimate a 1.10 release in April.

Please do not release any port based on a developmental build! Besides 
breaking compatibility you would cause problems due to unfinished 
changes. E.g. the current trunk has a few levels that are either 
unsolvable or can be shortcutted due to current unfinished reengineering 
work.

But it would be o.k. to apply the perfomance tuning patches to the 1.01 
release. Of course you would have to check for possible sideeffects. But 
I assume that the following patches should merge with 1.01 without 
problems:

Tunk revisions: 549-555, 557, 564, 566, 568, 569, 573

Till Harbaum wrote:
> Since i am also working on this, we should really try to keep things in sync.
> I am not a svn specialist. Can we ask the repository to create a branch for 
> the maemo port that inherits patches from the main branch and at the same 
> time dooesn't forward maemo specific patches upstream? That would imho be 
> perfect for pipeline and me.

We can either make a branch from the 1.01 release, apply the patches 
above and you can add any further Maemo patches you like. You would be 
free to release a 1.01 based Maemo version anytime. But of course we 
would need to merge the relevant changes lateron manually to the trunk.

Or we can make a working branch of the 1.10 trunk to prepare and test 
all Maemo changes necessary for a synchronous 1.10 release.

You would need Berlios developer account anyway. Please email me your 
account names for granting access to the Enigma repositories.

> If that's not possible i think we should collect a set of patches 
> (preferably "diff -Naur" or "patch" style, i still have to figure out how to 
> apply your diffs ...). The first one is the input patch pipeline just posted. 
> Other things that make sense for maemo (but also for other small/fixed screen 
> devices) is to remove/disable all unneeded resolutions and to disable the 
> installation of their graphics. This saves several megabytes disk space on 
> the target. E.g. maemo only needs 640x480 and the gfx32 tile set. Also maemo 
> needs some debian package stuff which imho might also go into the 
> repository/patch set. 

You should not remove anything within a Maemo branch. Stripping 
unnecessary stuff on the packaging stage should be sufficient.

I hope Tacvek and Erich will support you in aligning your patches in a 
mainstream mergable manner.

Greets,

Ronald